WELDING. 407 SEC. XIXc.—Influence of the metalloids upon welding.—The way in which, the impurities of the metal affect the welding power has been a matter of discussion, it having even been supposed that they act simply by interposition, and, again, that they increase the susceptibility of the iron to oxidation. I believe both of these theories are wrong. If the first were true, then one. per cent, of carbon would have the same effect as one per cent, of sulphur, which is manifestly not the case. The second theory does not hold, since sulphur, notoriously one of the worst enemies of welding, is not oxidized either in the acid Bessemer or open-hearth furnace, and there is no ground for assuming that it oxidizes in welding. As phosphorus, carbon and manganese protect iron from burning in the Bessemer and open-hearth, so they must also tend to be preferentially oxidized in a blacksmith's fire, and thus by preventing the formation of iron oxide, as well as by the formation of a liquid flux containing phosphoric acid and oxide of manganese, they should, as far as oxidation is concerned, assist rather than retard the welding. A third theory is that the impurities affect the mobility. When half of one per cent, of carbon is added to the metal, it produces a compactness or hardness, even when the steel is hot, that must prevent the easy flowing together that follows a pressure upon two pieces of white-hot wrought-iron or soft steel. A higher temperature cannot be used, because every increase in carbon reduces the safe working temperature at the same time that it increases the stiffness. This decrease in mobility doubtless plays an important part in the explanation, but I believe a greater influence is to be found in what may seem at first sight to be the same thing, but which is a different quality, viz.: The power, or property, of passing through a viscous state on the road to liquidity. Other metals, lead and copper for instance, are malleable and ductile, but do not go through a history of slow softening under the application of heat, the change to a liquid state being sudden and without any marked intermediate stage. Pig-iron is of the same character, for no matter how low the other metalloids may be, the presence of three per cent, of carbon produces a metal which changes suddenly from a solid to a liquid state, and it is reasonable to suppose that each increment of carbon, phosphorus and manganese tends in the same direction.